Torn between state and market: Private policy implementation and conflicting institutional logics^*

* An earlier version of this paper was presented at the workshop ‘The Causes and Consequences of Private Governance: The Changing Roles of State and Private Actors’ held on 6-7 November 2014 at the Mannheim Centre for European Social Research (MZES). I gratefully acknowledge funding for this event by the COST Action IS1309 ‘Innovations in Climate Governance: Sources, Patterns and Effects’ (INOGOV), MZES, and the Lorenz von Stein Foundation.

Abstract

Policy implementation by private actors constitutes a “missing link” for understanding the implications of private governance. This paper proposes and assesses an institutional logics framework that combines a top-down, policy design approach with a bottom-up, implementation perspective on discretion. We argue that the conflicting institutional logics of the state and the market, in combination with differing degrees of goal ambiguity, accountability and hybridity play a crucial role for output performance. These arguments are analyzed based on a secondary analysis of seven case studies of private and hybrid policy implementation in diverging contexts. We find that aligning private output performance with public interests is at least partly a question of policy design congruence: private implementing actors tend to perform deficiently when the conflicting logics of the state and the market combine with weak accountability mechanisms.

1 Introduction

This paper explores the conditions for the output performance of private policy implementers. The shift from government to governance (Bevir, 2011) has led to different forms of cooperation between the public and private sphere (Knill & Lehmkuhl, 2002). The inclusion of private actors and the introduction of market principles along the line of New Public Management (NPM) have considerably changed the way how public tasks, goods and services (outputs) are delivered (Hodge & Greve, 2007: 545–548). With the involvement of private actors the state has less direct control (Bevir, 2011: 459), but, with the exception of private self-regulation, ultimately remains responsible for safeguarding policy objectives (Knill & Lehmkuhl, 2002: 50). Evidence suggests that the (partial) privatization of public services is no guarantee for achieving policy outputs aligned with the public interest (e.g., Hodge & Greve, 2007: 553; Knill & Lehmkuhl, 2002; Schedler & Rüegg-Stürm, 2014: 46). Explaining how and why intended policy outputs are (not) delivered when private actors are delegated public tasks arguably presents a “missing link” for understanding the implications of private governance on policy implementation (Hodge & Greve, 2007: 545–546). In this paper we thus explore output performance, meaning the degree to which the output goals of a policy are met (Hupe & Hill, 2007: 294).

Scholars increasingly scrutinize the differing logics underlying public and private governance and their influence on implementing actors (Buffat, 2014; Considine & Lewis, 2003; Schedler & Rüegg-Stürm, 2014). This paper integrates the evidence that implementing actors are “embedded” within institutional contexts, which imply differing rationalities and identities. Hence, their actions are partly guided by underlying institutional logics (Garrow & Grusky, 2013; Thornton & Ocasio, 2008). Today's hybrid reality of policy implementation involves multiple, coexisting and potentially conflicting institutional logics, including an ideal-typical “state logic” focusing on legality, equity, security and correctness and a “market logic” focusing on profit, performance, competition, effectiveness and efficiency (Meyer, Egger-Peitler, Höllerer, & Hammerschmidt, 2014; Skelcher & Smith, 2014). Private and hybrid implementation modes entail a shift from actors predominantly adopting a state logic to also drawing on a market logic (Considine & Lewis, 2003; Fuenfschilling & Truffer, 2014: 9; Thornton & Ocasio, 2008).

We refer to conflicting logics as situations in which policy goals require an actor to draw on both logics and where the core values of the state logic contradict, are incompatible or incongruous with core values of the market logic (Skelcher & Smith, 2014: 441). Our central argument is that when the logics of the state and the market cannot be reconciled, then private actors tend to stick to the core values of the market that represent their institutional identity.

Private and hybrid implementation modes represent a set of government preferences for markets in goods and service delivery, and are assumed to represent the choice of the “best” implementation mode that “matches” the overarching policy goals (Howlett, 2009; Lytton & McAllister, 2014; Saetren, 2014). However, multiple policy design elements act in a complex interplay and can undermine one another in the pursuit of policy goals (Howlett, 2009; Howlett & Rayner, 2007). We thus formulate contrasting hypotheses on how three major design factors interact with conflicting logics on output performance, from both a top-down, “design” perspective (Howlett, 2009) and from a bottom-up, “implementation” standpoint (Saetren, 2014). These factors are (1) goal ambiguity (Chun & Rainey, 2005; Matland, 1995; Pandey & Wright, 2006); (2) accountability mechanisms (Klenk & Lieberherr, 2014); and (3) the implementation setting's hybridity (Knill & Lehmkuhl, 2002).

We illustrate our argument through a secondary analysis of seven case studies of private and hybrid policy implementation in diverging contexts. These studies cover both individual and organizational private implementing actors in (1) the Swiss food safety sector (Sager, Thomann, Zollinger, van der Heiden, & Mavrot, 2014), (2) the United States labour welfare sector (Dias & Maynard-Moody, 2007), (3) the Swiss telecoms sector (Ingold & Varone, 2014) and (4) the English water sector (Lieberherr, 2012). By distinguishing different types of goals according to whether they follow a market or a state logic, we find that the conflicting logics of the state and the market combined with weak accountability help to understand why private implementing actors may not achieve policy outputs as intended. We can thus tentatively specify conditions under which private and hybrid implementation modes can be incongruent with a policy's targets (Howlett, 2009: 83).

Next we discuss the involvement of private actors in policy implementation and then turn to the link between institutional logics and output performance in relation to ambiguity, accountability, and hybridity. In section three we introduce the research design. After presenting our findings in section four, section five provides a comparative discussion and the concluding section puts them into perspective.

2 Private and hybrid implementation and output performance

We address two implementation modes, where private actors deliver public measures, goods and services (hereafter referred to as output) (Bevir, 2011: 467). First, in hybrid modes, private actors participate in implementation alongside public actors (excluding accountability mechanisms). Second, in private modes only private actors carry out implementation tasks. The latter are for-profit actors (as opposed to citizens or non-profit organizations) which are either individuals or organizations (Winter, 2003: 219).

2.1 Potentially conflicting logics of the state and the market

Institutional logics denote ideal-typical “socially constructed, historical patterns of cultural symbols and material practices, including assumption, values, and beliefs, by which individuals and organizations provide meaning to their daily activity, organize time and space, and reproduce their lives and experiences” (Thornton, Ocasio, & Lounsbury, 2012: 2). By guiding and constraining both organizational and individual behaviour, institutional logics impact output performance (Garrow & Grusky, 2013; Skelcher & Smith, 2014: 438). Table 1 outlines the characteristics of the ideal-typical state and market logics. Actors are usually embedded within several institutional sectors, each of which exhibits a distinct logic (Fuenfschilling & Truffer, 2014: 8–9; Thornton & Ocasio, 2008: 111).³

Table 1 Ideal-typical institutional logics of the state and the market.

	State	Market
Source of authority	Bureaucracy, government	Shareholders
Source of legitimacy	Democratic participation; procedural	Share price; results-based
Mode of governance	Bureaucratic governance based on laws, rules, directives	Contractual governance based on objectives/targets, results, performance measures, and management tools in a competitive environment
	Accountability towards the sovereign	Accountability towards shareholders & customers
Vision and mission	State as sovereign, serving society and the public interest	Achieving objectives/targets and serving customers
Basis of norms and attention	Citizenship, status in interest group	Self-interest, status in market
Core values	Legality, equity, security, correctness	Performance, effectiveness, competition and efficiency
Basis of strategy	Increase community good	Increase efficiency, profit

Sources: Own representation based on Meyer et al. (2014: 6), Fuenfschilling and Truffer (2014: 787), Skelcher and Smith (2014: 6).

Private actors are traditionally embedded in market institutions thus and follow a “market logic” according to values of performance, efficiency, competition and profitmaking. Simultaneously, workers in the private sector now represent public policy to the people. Thus, private implementing actors become exposed to the “state logic”, which implies a public service rationality that focuses on such values as legality, equity, security, correctness needed to provide public services (Skelcher & Smith, 2014: 439). Additionally, implementation by private actors is often motivated by goals like efficiency and competitiveness, which are then defined in, e.g., contracts. The public-private dichotomy thus becomes increasingly obsolete (Knill & Lehmkuhl, 2002).

Under hybrid and private implementation, conflicting logics exist when the private implementing actor is unable to reconcile tensions between logics, i.e. catering to the core values of the state cannot be reconciled with market values, or vice versa. Such incompatibilities can occur even if both types of values are contained in goals prescribed by a policy, contract or the like (Howlett & Rayner, 2007). For instance, if the private actors are unable to make a profit while delivering the public service equitably to all segments of the population, then there would be a conflict (Lieberherr, 2012).

As private implementing actors’ “actions, intentions, and interests are themselves institutionally conditioned in systematic ways” (Garrow & Grusky, 2013: 104), their response to conflicting logics is linked to their organizational identity in the for-profit realm (Skelcher & Smith, 2014: 446). Our core expectation is that they keep and, in case of conflict, prioritize an entrepreneurial logic:

If a public policy draws on the logics of the state and the market and these are in part or fully in conflict, then the private implementing actor tends to prioritize the market logic over the state logic; output performance thus tends to be deficient.

However, in reality conflicting logics never occur in a vacuum. The question arises how other design factors interact with conflicting logics on output performance.

2.2 Hypotheses

The output performance of implementing actors depends upon their degree of autonomy from political authorities (Lipsky, 1980). Discretion is shaped by policy design, different forms of managerial or social control, and accountability (Howlett, 2009; Hupe & Hill, 2007).

Top-down and bottom-up implementation perspectives hold contradictory views of the role and effects of discretion (Matland, 1995). Top-down perspectives assume a direct link between policy design and policy implementation. If goals and implementation modes are well calibrated and match (“congruence”), then policy targets can be achieved (Howlett, 2009: 74, 82; May, Sapotichne, & Workman, 2006). When discretion is high, it is easier for private implementing agents to deviate from output goals (e.g., Knill & Lehmkuhl, 2002). Conversely, bottom-up implementation perspectives emphasize that implementing actors do not just respond to rules, but use their discretion to adjust them (Winter, 2003). Discretion thus helps implementers to be politically and contextually responsive, to correct for policy failures and fulfil their tasks as appropriate (Chun & Rainey, 2005: 537; Lipsky, 1980).

To integrate both perspectives (Saetren, 2014), we hereafter present three design factors that intervene with conflicting logics by shaping discretion and its use. We hypothesize the diverging directions of these interactions from a top-down and a bottom-up perspective.

2.3 Ambiguity

Goal ambiguity refers to the lack of clearly defined goals by a policy, i.e. the extent to which a set of goals allows leeway for interpretation (Chun & Rainey, 2005: 531).

From a top-down perspective, ambiguity can create uncertainty and policy incoherence (Chun & Rainey, 2005: 537; May et al., 2006): “as the clarity and precision of goals diminishes, it becomes […] likely that policy means and ends will be mismatched” (Howlett, 2004: 5). By enhancing the complexity of the environment, policies with conflicting institutional logics can reinforce ambiguity, and vice versa (Pandey & Wright, 2006).

H1a

The conflicting logics of the state and the market in combination with high ambiguity tend to negatively affect output performance, while their combination with low ambiguity tends to result in appropriate performance.

Conversely, from a bottom-up perspective, “the clearer goals are the more likely they are to lead to conflict” (Matland, 1995: 158). Goal clarity hinders the implementing actors in reacting to the challenges arising during implementation. In contrast, goal ambiguity broadens the room for interpreting appropriate goals, learning, and adjustments which are needed to avoid conflict and maintain policy coherence (Chun & Rainey, 2005: 537; May et al., 2006):

H1b

The conflicting logics of the state and the market in combination with high ambiguity tend to result in appropriate output performance, while their combination with low ambiguity tends to have a negative effect on performance.

2.4 Accountability mechanisms

The implementing actor's discretion can be controlled through various accountability mechanisms (Klenk & Lieberherr, 2014), defined as the assignment of responsibility to certain public and private actors to directly or indirectly enforce policies vis-à-vis the private implementers.

From a top-down view, accountability mechanisms can ensure that the public rationale is not overridden by the private rationale, as private implementing actors are held to account to a public actor (Knill & Lehmkuhl, 2002). Deviations from output goals will be detected and have negative consequences (Howlett, 2009: 81; Lytton & McAllister, 2014):

H2a

The conflicting logics of the state and the market in combination with strong accountability mechanisms tend to result in appropriate output performance, while their combination with weak accountability mechanisms tends to have a negative effect on performance.

Conversely, bottom-up views stress the simultaneous coexistence of multiple demands for accountable behaviour for private implementers, which can pose inescapable dilemmas (Hupe & Hill, 2007) and hence negatively affect output performance (Buffat, 2014; Lipsky, 1980). Strong public-administrative accountability mechanisms can reinforce the conflict of logics, and limit the discretionary capacity of the private implementing actors to mediate this conflict:

H2b

The conflicting logics of the state and the market in combination with strong accountability mechanisms tend to negatively affect output performance, while their combination with weak accountability mechanisms tends to result in appropriate performance.

2.5 Hybridity

Whether the implementation mode is private or hybrid matters. Hybridity entails mutual relationships between public and private actors, with a plurality of institutional logics giving implementing actors multiple identities (Skelcher & Smith, 2014).

From a top-down perspective, mechanisms that encourage cooperation foster the reconciliation of public and economic interests (Knill & Lehmkuhl, 2002: 46; Pandey & Wright, 2006: 525). The regular dialogue between private and public actors facilitates control over task execution (Lytton & McAllister, 2014: 331). The logics of the state and the market can hence better be reconciled if they coexist, rather than the latter being dominant:

H3a

The conflicting logics of the state and the market in a hybrid implementation setting tend to result in appropriate output performance, while in the absence of hybridity, they tend to have a negative effect on performance.

In contrast, bottom-up scholars would argue that hybridity multiplies the complexity of interrelations and the potentially contradictory accountabilities of the private implementing actors (Buffat, 2014: 84), which reduces the latter's capacity to reconcile conflicting demands. Mutual interdependencies between public and private actors can hamper, rather than enhance the formers’ ability to hold the latter accountable (Sager et al., 2014):

H3b

The conflicting logics of the state and the market in a hybrid implementation setting tend to negatively affect output performance, while in the absence of hybridity, they tend to result in appropriate performance.

Fig. 1 illustrates our framework. The theoretical underpinnings of the institutional logics approach are shaded grey, whereas the variables and hypotheses are shaded black. No conflict of logics prevails when the logics of the state and the market cohere, i.e., overlap; otherwise, private actors may have to opt for one of the logics.

Fig. 1 Conflicting logics framework. Source: Own representation. Note: the order of the design factors does not indicate a temporal or causal sequence. Bold: Top-down hypotheses, italics: bottom-up hypotheses.

Ambiguity, accountability mechanisms and hybridity also interact with each other; however, this is not our focus. In addition, more possible explanations for output performance exist. We retain a certain explorative character in our assessment to detect such patterns.

3 Data and methods

3.1 Measurement

Since institutional logics are difficult to measure (Thornton & Ocasio, 2008), this deserves special attention (see Table 2 ). We emphasize material over cultural components of institutional logics, i.e., norms, attention as well as resulting core values and strategies outlined in Table 1 (Skelcher & Smith, 2014: 440). Referring to rule-based behaviour, we address the goal orientations, outcomes, rules and appropriate means implied by either the state or the market that are supposed to guide the private actors’ behaviour (Garrow & Grusky, 2013).

Table 2 Concepts and measurement.

Variable	Measurement
Logic of the state	Those behaviours and outcomes that the private actors need to pursue in order to achieve legality, equity, security, correctness. Measures, goods and services that should be delivered according to the policy, mandate, contract or task delegating document (=output goals), which cater to e.g. equitable service provision and security of supply.
Logic of the market	Those behaviours and outcomes that the private actors need to pursue in order to generate rate-of-return, remain competitive, enhance performance, efficiency or not alienate shareholders or customers. Measures, goods and services that should be delivered according to the policy, mandate, contract or task delegating document (=output goals), which aim at enhancing performance, efficiency, profit, and competition.
Conflict of logics	Comparison of state and market logics. Yes = at least some contradictions between logics (achievement of core values of state logic at least partially only at cost of core values of market logic and vice versa). No = no contradictions (can achieve both core values without negatively affecting the other).
Goal and task ambiguity	High = at least partially lacking definition of output goals, tools for implementation and their use, and roles of organizations involved in implementation. Low = existence and clarity of definition and/or clarity of output goals, tools for implementation and their use, and roles of organizations involved in implementation.
Accountability mechanisms	Strong = effective regulation or enforcement/oversight mechanisms at the system, organizational or individual level; if implementing actors act and decide autonomously against the policy goals, then it will be detected or they suffer from unfavourable consequences. Weak = at least partial absence of effective regulation or enforcement/oversight mechanisms at the system, organizational or individual level: implementing actors have much freedom to act and decide autonomously without having it detected or suffering from unfavourable consequences.
Hybridity	Yes = public and private actors collaborate in implementation tasks (output delivery as defined in the policy, mandate, contract or task delegating document; including enforcement tasks vis-à-vis target groups, but excluding regulation & oversight of output delivery). No = only private actors involved (no collaboration).
Output performance	Comparison of output goals with actually delivered output (assess if delivered outputs are in line with the specified goals): Appropriate = actually delivered output coheres with output goals. Deficient = at least partially lacking coherence between deliverance and intended goals.

The logic of the state is measured through core values such as legality, equity, security, correctness. These values can translate into the policy's output goals: the measures, goods and services that should be delivered according to the policy, mandate, contract or task delegating document which aim at the core values of the state logic: e.g., ensuring equitable service provision and security of supply. We can equally observe the state logic through those behaviours and outcomes that the private actors need to pursue in order to realize the core values of the state. The logic of the market is measured through core values such as performance, profit, competition and efficiency. These can manifest themselves as those outcomes that the private actors need to pursue in order to generate a rate-of-return, remain competitive, enhance efficiency or not alienate shareholders or customers; as well as output goals which aim at the same ends.

We do not measure the state and market logic through actual performance, but through statements on what behaviour would be in line with the core values outlined in Table 1. As mentioned above, an actor may draw on both state and market logics. Identifying conflicting logics of the state and the market then requires a comparison of the core values and related output goals of these two logics to find contradictions between them.

3.1.1 Case selection

To illustrate our arguments, we conduct a secondary analysis of four published empirical studies. We have purposively selected empirical examples that fulfil three preconditions to be comparable: private actors deliver output tasks; output goals are defined to a certain degree; and the studies contain information about all the variables of interest (Blatter & Blume, 2008: 336, 341). Since the research design and analysis overlap (Gerring, 2008), we define our cases such that they are both similar and different enough to treat them as comparable instances of the relevant phenomena. The unit of variation for our argument is whether the goals draw more on the logic of the state or of the market. The compatibility of this main logic with the actors’ dominant logic translates into differences in performance. We hence conceive of our cases as types of policy goals which differ in the main logic that they follow. These cases are “embedded” within the respective countries and policy sectors, some drawing on the same legal arrangements. Our comparison then implies a combination of three case selection strategies (Gerring, 2008: 677).

First, the proposed causal mechanism between conflicting logics and output performance is best supported, if we can demonstrate the robustness of the relationship across contrasting contextual settings, i.e. within different sectors (Hupe & Hill, 2007: 293), institutional settings and at two levels of analysis (Winter, 2003). Second, our casing strategy enables us to consider both appropriate performance, i.e. types of output goals which were met, and deficient performance where this was not the case. The combination of a most different systems design with the method of difference facilitates a clear logic of elimination when comparing the cases to identify the relevant causal factors (Levi-Faur, 2006). Third, our cases are “diverse” regarding the three design factors which we assume to interact with conflicting logics (Gerring, 2008: 650 ff).

These criteria lead us to focus on seven cases (types of policy goals) which are nested within four different policy areas and three countries (Table 3 ): the United States (US), United Kingdom (UK), and Switzerland (CH).

Table 3 Case selection.

	Hybrid mode	Private mode
Individual level	Food safety CH Case 1: Output goals Sager et al. (2014)	Labour welfare US Case 2: Job placement goals Case 3: Case management goals Dias and Maynard-Moody (2007)
Organizational level	Telecoms CH Case 4: Common good goals Case 5: Economic goals Ingold and Varone (2014)	Water supply UK Case 6: Economic goals Case 7: Common good goals Lieberherr (2012)

Source: Own representation.

We adopted several strategies to address possible methodological pitfalls of secondary analyses (Blatter & Blume, 2008). Especially in the absence of close knowledge of the primary data underlying cases 2 and 3, there is always a way to construct a story coherent with the expectations (“storytelling”). To allow for alternative explanations, we have formulated rivalling expectations for each factor: it is impossible to simultaneously confirm contrasting hypotheses (Blatter & Blume, 2008). To further avoid a confirmation bias, our cases studies remain as faithful as possible to the original narrative of the published studies. Yet our secondary analysis goes beyond a narrative review by interpreting the material using our conflicting logics framework, see Fig. 1.

4 Analysis

We first conduct in-depth within-case analyses in light of the variables of interest and their interaction (Table 2 and Fig. 1). In a second step, the cases are first discussed and then compared using Levi-Faur's (2006) inferential strategy.

4.1 Food safety in Switzerland (case 1)

Sager et al. (2014) have analyzed the hybrid implementation of the Swiss Ordinance on Veterinary Medicinal Products (OVMP). Alongside the cantonal public inspectors (public veterinarians), the private veterinarians coproduced outputs, being legally responsible for ensuring livestock famers’ compliance with the OVMP, who buy the former's services.

The output goals were unambiguous: the private veterinarians concluded written agreements with their customers — the livestock farmers — which entailed the duty to perform biannual visits on the farms to check compliance with the OVMP. The farmers paid the veterinarians for these visits. Furthermore, each manufacture of so-called medicated feedingstuffs by farmers on the agricultural site had to be supervised by the veterinarian. Accountability mechanisms of these private veterinarians were virtually absent: the official controls focussed only on their role as policy addressees, and the use of sanctions against the private veterinarians was uncommon. There was no national database for the official controls of the private veterinarians, and hence, no federal oversight.

The business relationship between the private veterinarians and the livestock owners created specific market-oriented goal orientations: it was important “for the private veterinarians not to annoy their customers. Most [livestock farmers] clearly prefer collaborating with a veterinarian who does not emphasize compliance too much. (…) Moreover, private veterinarians are locally rooted, and a reputation of being ‘strict’ could potentially alienate other customers (…) the easiest way for veterinarians to avoid a loss of customers is if they can bypass unfavourable provisions without having it discovered (…) private veterinarians have a significant economic interest in not carrying out their duties as implementing actors, and the interdependencies created by the business relationship have the potential to aggravate this effect” (Sager et al., 2014: 497). The logics of the state and the market are contradictory because an “effective enforcement of the OVMP's regulations is a potential threat to the veterinarians’ business” (Sager et al., 2014: 497). As service providers, the private veterinarians were economically dependent upon their customers, which limited their capability to effectively act as enforcement actors (Sager et al., 2014: 498).

Sager et al. (2014) explain the output performance of the private veterinarians accordingly: the latter are not primarily implementing actors, but service providers, which is why economic motivations outweigh the responsibility as implementing actors for many private veterinarians (Sager et al., 2014: 397–398). Biannual visits are carried out only in half of the cases and data indicate that a great part of the private veterinarians substantially neglect their obligation to ensure the compliance of farmers with the OVMP's regulations.

4.2 Labour welfare in the United States (cases 2 and 3)

Dias and Maynard-Moody (2007) have examined the performance of caseworkers in the fully privatized for-profit welfare-to-work training programme WorkOpts serving long-term recipients in Porter City, U.S. (anonymized study). This was implemented in the course of a welfare reform programme that encouraged states to contract with for-profit firms.

4.2.1 Job placement goals

The contract had highly unambiguous performance-based job placement goals (case 2): it required WorkOpts to serve at least 1200 clients per year, to place ten per cent of them in the workforce for at least 90 days, and to provide recipients with 30 h of work activities. The necessity to renew the contract each year, bound to clear goals and financial rewards, effectively limited the discretion of WorkOpts managers to deliver on these job placement goals (Dias & Maynard-Moody, 2007: 199). Managers thus exercised strong managerial oversight over frontline staff to dedicate their time to the financially rewarded contract goals.

The job placement goals drew on a market logic and required WorkOpts to meet the city's financially rewarded job placement goals and the mother company's profit requirement of at least eight per cent (Dias & Maynard-Moody, 2007: 195, 200). This implied maximizing efficiency: “the easiest way for WorkOpts’ management to meet the city's contract goals and make profit was to minimize the time and effort line staff devoted to each client” (Dias & Maynard-Moody, 2007: 198).

The financially motivated managers “chose a narrow reading of the contract [which] meant fully adopting the work-first model of practice at the expense of the (…) case management model” (Dias & Maynard-Moody, 2007: 200).

4.2.2 Case management goals

The contract also stipulated more ambiguous case management goals (case 3) entailing services for enrolled clients (e.g., individual assessment, job coaching etc.), however without attaching financial inducements or indicating clear amounts of time to be invested. Caseload size was not to exceed 35 clients per case manager (Dias & Maynard-Moody, 2007: 194–195, 200). The contract hence granted much flexibility and little monitoring and enforcement (Dias & Maynard-Moody, 2007: 201).

The case management goals drew on a state logic, but “distracted from the goal of immediate job placement, which compromised the parent company's economic gain” (Dias & Maynard-Moody, 2007: 198). Frontline workers were urged to adopt a “rush-rush” style with clients. This was incompatible with their own case management priorities (Dias & Maynard-Moody, 2007: 201–2). This resulted in an “inability to reach some form of reconciliation between these two perspectives” (Dias & Maynard-Moody, 2007: 204).

Consequently, output was unsatisfactory. “Although in theory the programme could feasibly implement both mandates, management had little incentives to do both” (Dias & Maynard-Moody, 2007: 207). Case managers had much higher caseloads than prescribed (70–100, prescribed: 35), and only about ten minutes to deal with each client.

4.3 Swiss Telecommunications (cases 4 and 5)

Ingold and Varone (2014; see also Ingold, Varone, & Stokman, 2013) have analyzed the implementation of the Swiss telecommunications act. In 1997 a major legal reform led to partial privatization of the former monopolist and defined rights and duties of the newly created company Swisscom, the majority still in public hand. Within the fragmented hybrid setting, public actors such as the Swiss Communications agency OFCOM or the sector-specific regulator ComCom were involved in universal service delivery. Swisscom can only be held accountable for its actions ex post. Two types of public policy goals of telecommunications regulations and universal service delivery were outlined in an unambiguous, detailed and transparent way.

4.3.1 Common good goals

Common good goals (case 4) drawing on a state logic concerned current technological standards, the needs of the consumers, but also the universal access guarantee for each Swiss household to fixed telephony infrastructure until 2017 and non-discrimination of handicapped people. Quality standards and upper price limits were defined by the government. By 2009, quality and technology reached very high standards, and universal service provision was achieved (Evaluation by the Federal Council 2009).

4.3.2 Economic goals

Economic goals (case 5) drawing on a market logic expected Swisscom to cover all costs. In the vein of the full liberalization of the Swiss telecoms sector, Swisscom was furthermore legally obligated to comply with free market rules and grant open access to households to all other competing operators. The access to all households through the universal service delivery mandate created a market advantage for Swisscom, and the incentive to exploit this situation by denying free-access to all other operators. This advantage fully conformed to the maximization of Swisscom's self-interest. The state logic (universal service delivery) thus conflicted with the market logic (liberalization). Swisscom differed from the other private implementing actors in facing the challenge to adopt not only the common good goals but also the economic goals.

Different judicial disputes involving Swisscom, regulators and courts reveal that Swisscom seemed to have difficulties in fully adopting a core value of the market, namely, competition. Rather, Swisscom tended to exploit its market position inherited from its former institutional identity as a public monopolist.

4.4 English water supply (cases 6 and 7)

Lieberherr (2012) analyzed the performance of the for-profit water provider Yorkshire Water, who is authorized by the 1989 Water Act to provide drinking water and sanitation (collection and treatment of wastewater) services in Yorkshire County, UK. The national government fully divested the ownership of the infrastructure and the responsibility to provide water supply and sanitation services to private actors. Two types of clearly formulated policy goals can be identified in the relevant water acts.

4.4.1 Economic goals

Economic goals (case 6) aimed at increasing efficiency through performance measures and attracting private capital (Lieberherr, 2012). The policy's economic goals drew on a market logic. The private actor had to follow a for-profit strategy in order to remain attractive for its international investors. Yorkshire Water's discretion was limited due to the government's extensive system of arms-length regulation. In the 1990s and early 2000s, the water company became highly regulated. Under this “performance-based regulation, (…) Ofwat [the economic regulator] ha[d] the responsibility for the final decision, it ha[d] ‘the trump card”’ (Lieberherr, 2012: 192). Accountability mechanisms were strong for ensuring economic goals. Yorkshire Water had a “cost-plus tariff system (…) that generate[ed]s a return on capital for investors (…) Yorkshire Water [was] obligated to transfer revenue to its parent company, which then distribute[d] this in the form of dividends to the investors” (Lieberherr, 2012: 148). This system implied a profit-maximizing strategy (Lieberherr, 2012: 171).

The company fulfilled the economic goals: “Yorkshire Water (…) surpassed Ofwat's [the economic regulator's] efficiency targets from 2005–2010, thus becoming ‘the benchmark in the latest comparative competition”’ (Lieberherr, 2012: 201). The company was also successful at attracting private capital, where “the shareholders have consistently received dividends (…) average rate of return on capital at Yorkshire Water [was] ca. 24%” (Lieberherr, 2012: 206).

4.4.2 Common good goals

Common good goals (case 7) drawing on a state logic specified universal service provision (aim of 100% household connection rate) and affordability (citizens should spend less than 3% of their disposable income on water services) (Lieberherr, 2012: 91). Accountability mechanisms were weak in safeguarding common good goals, as the economic regulator focused on efficiency (Lieberherr, 2012). The state retained no managerial control, as the operations and ownership rights were 100% in private hands (Lieberherr, 2012: 147). Moreover, the company had “to respond to what the shareholders want” (Lieberherr, 2012: 171). Since the common good goals of affordability stood at odds with the economic goals of driving efficiency and attracting private capital, the logics of the state and the market conflicted.

Consequently, the operator pursued economic goals partly at the cost of the common good goals. The company performed well in terms of household connections, as 100% of the household were connected to water supply and 97% to sanitation in 2011 (Lieberherr, 2012: 191). However, operator was unable to lower or stabilize water prices, which continued to rise above the rate of inflation and water poverty exists: “households spending over 3% of their disposable income on WSS [water supply and sanitation] services” (Lieberherr, 2012: 91).

5 Discussion and comparison

Our case studies have illuminated the causal mechanisms that link conflicting institutional logics of the state and the market with private implementing actors’ output performance, as well as their interplay with other design factors.

5.1 Discussion

In the case of private food safety inspectors in Switzerland (case 1), Conflicting logics played a crucial role for their deficient output performance. Goal clarity made it impossible for the private veterinarians to satisfy both their customers and the policy goals (H1b). The weak accountability mechanisms made it easy to deviate from the output goals (H2a). Conversely, no direct effect of hybridity on conflicting logics or output performance could be detected (H3a or b not evidenced). However, hybridity created complex interrelations which further weakened accountability mechanisms: as public and private veterinarians are professional peers, the former were reluctant to sanction the latter (Sager et al., 2014: 495).

That output goals were (in)compatible with the market logic proved decisive for output performance of the private implementers of labour welfare services in the United States (cases 2 and 3). In line with H1a, the high ambiguity of the case management goals (case 3) contributed to the management's prioritization of the unambiguous job placement goals (case 2) at the expense of the former. The vague case management goals (case 3) were simply less decisive for contract renewal, as goal ambiguity contributed to the absence of stringent accountability mechanisms, which reinforced the negative effect of conflicting logics in case 3 (H2a). Strong accountability mechanisms helped ensure appropriate performance in case 2. The dominance of the market logic in the non-hybrid setting implied a complete absence of legal and financial support for proponents of the state logic (case 3). While this lowered accountability, low hybridity did not impact conflicting logics or performance neither in case 2 nor 3 (H3a or b not evidenced).

Conflicting logics were also crucial for output performance in the Swiss Telecommunications sector (cases 4 and 5). Ambiguity did not have an observable causal impact on output performance (no evidence for H1a or b). In case 5, low and ex-post accountability contributed to the negative impact of the conflicting logics between the market and the state on economic output performance (H2a). Ex-post accountability, however, seemed to have no negative impact on public output performance when the logics were not conflicting (case 4). Finally, hybridity neither influenced conflicting logics nor output performance (H3a or b not evidenced). But hybridity reinforced the fragmented regulatory arrangements and thus the shortcomings of the weak ex-post accountability structures, allowing regulators and courts only to intervene when deficient outputs are already present, impacting mostly deficient output performance in case 5.

Conflicting logics equally played a key role for output performance of the analyzed private English water supplier (cases 6 and 7). However, no clear causality could be found between ambiguity and output performance (no evidence of H1a or b). The strong accountability mechanism enabled the private actor to pursue economic goals, in line with the interests of the investors (case 6) and fostered (to a degree) by the system of economic regulation which integrated a market logic, but at the cost of certain common good goals where accountability mechanisms were weak (case 7) (H2a). Finally, case 7 indicates a link between a lack of hybridity, conflicting logics and weak accountability means for deficient performance (H3a). Due to the full divestiture the private company had high discretion; along with the integration of the market logic, this led to a deficient fulfilment of common good goals. Conversely, a link between no hybridity, coupled with strong accountability and the absence of conflicting logics, and appropriate output performance (case 6) was found.

Overall, the analyzed implementing actors have consistently reacted to conflicting logics following their institutional identity as private or, in the case of Swisscom, formerly public actors. Our results further show that if logics cannot be reconciled, then this does not mean that overall goal attainment completely fails.

Based on Table 4 , we now compare our cases to see which causal factors have a robust and consistent effect across different contexts (policies, countries, and levels of aggregation). By maximizing variance on both the dependent and “control” variables, we can “eliminate the variables that are less likely to exercise a causal effect on the different outcomes since they appear in both cases” (Levi-Faur, 2006: 60). Applying this strategy, our results suggest that private implementing actors tend to perform deficiently when the logics of the state and the market conflict, whereas in the absence of conflict, output goals tend to be achieved.

Table 4 Cross-case comparison.

Case	Defined output goals	Conflicting logics	Ambiguity	Accountability	Hybridity	Performance
Food Safety CH Case 1	Output goals	Yes	Low	Weak	Yes	Deficient
Welfare US Case 2	Job placement goals	No	Low	Strong	No	Appropriate
Welfare US Case 3	Case management goals	Yes	High	Weak	No	Deficient
Swisscom Case 4	Common good goals	No	Low	Weak	Yes	Appropriate
Swisscom Case 5	Economic goals	Yes	Low	Weak	Yes	Deficient
Water UK Case 6	Economic goals	No	Low	Strong	No	Appropriate
Water UK Case 7	Common good goals	Yes	Low	Weak	No	Deficient

We further find that weak accountability mechanisms consistently reinforce the negative effect of conflicting logics on output performance. The top-down view of our Hypothesis 2a hence finds support. In the absence of conflicting logics, however, output performance can still be appropriate even if accountability mechanisms are weak. Furthermore, our data do not allow us to conclude that stronger accountability mechanisms would resolve the problem — simply because conflicting logics never occurred in combination with strong accountability mechanisms.

Evidence is not consistent regarding other design factors as hypothesized. While both goal ambiguity and clarity can contribute to conflicting logics and deficient output performance (Chun & Rainey, 2005; Pandey & Wright, 2006), it did not play a causal role in two of the four policy sectors. We find no evidence that the hybridity of implementation modes consistently affects output performance (Knill & Lehmkuhl, 2002). However, these design factors do matter, as they mutually impact each other and specifically accountability. Both the absence and the presence of hybridity can further weaken accountability mechanisms on implementing actors. The same holds for high goal ambiguity in the US case. The precise role of goal ambiguity and hybridity and the latter's interplay with accountability (Buffat, 2014) should be subject to future research.

6 Conclusion

This paper investigated the role of institutional logics and different design factors for the output performance of private implementers of public policies (Thornton & Ocasio, 2008; Schedler & Rüegg-Stürm, 2014). We have illustrated how such private or hybrid modes of implementation involve two differing logics: a logic of the state, characterized by values as legality, equity, security and correctness; and a logic of the market, following the core values of performance, efficiency, competition and profit-oriented principles (Meyer et al., 2014). Implementing actors are required to draw on multiple coexisting logics, which may complement each other, but they can also clash (Skelcher & Smith, 2014).

Our secondary analysis of seven cases embedded in three different countries and four policy sectors suggests that the conflicting institutional logics of the state and the market can play an important role for understanding why the performance of private implementers may or may not conform to policy goals (Hodge & Greve, 2007). If the private implementers cannot reconcile the state logic with their own market logic, then they tend to prioritize the latter. Swisscom, as a formerly public actor, prioritized the state logic. Weak accountability mechanisms seem to reinforce this negative effect (Klenk & Lieberherr, 2014; Lytton & McAllister, 2014). Our results also illustrate that both top-down and bottom-up perspectives on the role of discretion can be valid under certain circumstances (Saetren, 2014). However, accountability mechanisms remain crucial for control from the top (Klenk & Lieberherr, 2014; Lytton & McAllister, 2014). Aligning private output performance with public interests is thus at least partly a question of policy design congruence (Howlett, 2009; Lytton & McAllister, 2014; May et al., 2006).

For scholars of private governance, our study illustrates the usefulness of the institutional logics approach for gaining a profound understanding of policy implementation. When integrating the notion of conflicting institutional logics, the responses of private policy implementers become partly predictable. For policymakers, this suggests that at least to a degree, potential conflicts between goals of the state and the market can and should be anticipated ex ante and, if possible, avoided. Otherwise, core values of the state such as equitable service provision and security of supply might be overridden by market-oriented goals (Howlett & Rayner, 2007).

This study has scrutinized a purposive set of cases which fulfilled certain prerequisites to analyze our argument. We have drawn conclusions from our observations regarding the accuracy and consistency of our theory, also known as “analytic” rather than statistical generalization (Blatter & Blume, 2008: 341). Future research should sensibly test our conflicting logics argument in other cases, based on primary data specifically collected for this purpose. Due to our focus on robust effects across contexts, we also did not account for complex interactions between different policy design elements (Howlett, 2009). Issues of goal ambiguity and accountability mechanisms are, for instance, closely intertwined (Matland, 1995). Notwithstanding these limitations, in linking the institutional logics framework with a policy design and implementation perspective, we have identified the conflicting logics of the state and the market as a specific condition under which output performance might suffer in private policy implementation. This study thus sheds light onto what constitutes a “missing link” for understanding the implications of private governance.

Notes

* An earlier version of this paper was presented at the workshop ‘The Causes and Consequences of Private Governance: The Changing Roles of State and Private Actors’ held on 6-7 November 2014 at the Mannheim Centre for European Social Research (MZES). I gratefully acknowledge funding for this event by the COST Action IS1309 ‘Innovations in Climate Governance: Sources, Patterns and Effects’ (INOGOV), MZES, and the Lorenz von Stein Foundation.

3 We only elaborate on the relationship between the state and the market logic.